Azure Function Plans and Hosting Options

When it comes to Azure Functions, you have several hosting options to choose from, but did you know that the Consumption Plan is the most cost-effective option? It's ideal for small-scale, event-driven workloads.

Azure Functions also offers the Premium Plan, which provides dedicated resources and is suitable for applications that require a high level of isolation and predictability. This plan is perfect for large-scale, long-running workloads.

The App Service Plan is another option, which allows you to host Azure Functions alongside other web applications, providing a more traditional hosting experience. This plan is ideal for applications that require a high degree of customization and control.

Each hosting option has its own pricing model, with the Consumption Plan charging per execution, the Premium Plan charging per instance, and the App Service Plan charging per virtual machine.

Readers also liked: Azure App Service Plan

Azure Function plans offer a range of options to suit different needs and scenarios. You can choose from Consumption, Azure App Service, and Azure Functions Premium plans.

The Consumption plan is a cost-effective option, where you're billed based on function executions. It's ideal for small-scale applications or those with intermittent usage.

Azure App Service plans are suitable for users who already utilize App Service to run mobile, API-based, and web applications. They can run Azure Functions at no additional cost.

The Azure Functions Premium plan offers enhanced performance, VNET access, and no cold start or execution charge. However, it has the same features as the Consumption plan.

Here's a summary of the plans:

Azure Function plans also have specific timeout durations, which vary depending on the plan. For example, the Consumption plan has a default timeout of 5 minutes, while the Premium plan has an unlimited timeout duration.

See what others are reading: How to Increase Azure Function Timeout

If you're looking for a dedicated hosting plan for your Azure Functions, you have four options: Basic, Standard, Premium, and Isolated SKUs.

These dedicated App Service plans are analogous to a server farm in conventional hosting, providing a set of dedicated compute resources for your app to run.

Readers also liked: Windows Azure Hosting Plans

You can configure one or more function apps to run on the same computing resources as other App Service apps, such as web apps.

Free and Shared tier App Service plans aren't supported by Azure Functions, so you'll need to consider alternative options like the Consumption plan or the Flex Consumption plan, which are billed based on function executions.

Consider a dedicated App Service plan if you have existing, underutilized VMs that are already running other App Service instances.

You can also use a dedicated App Service plan if you want to provide a custom image on which to run your functions.

Here are the dedicated App Service plans supported for function app hosting:

Always On is a crucial setting to enable when running your app on an App Service plan. This setting ensures your function app runs correctly, even after a few minutes of inactivity.

If you choose not to enable Always On, you can reactivate an idled app in two ways: send a request to an HTTP trigger endpoint or any other endpoint on the app, or access your app in the Azure portal. Even a failed request can wake up your app.

The execution timeout for individual functions is controlled by the functionTimeout setting in the host.json project file. This setting is separate from the Always On setting.

To run JavaScript (Node.js) functions on an App Service plan, choose a plan with fewer vCPUs. This is recommended for optimal performance.

Expand your knowledge: Azure Move Web App to Another Service Plan

Running in an App Service Environment lets you fully isolate your functions and take advantage of higher numbers of instances than an App Service Plan.

To get started, you need to see Introduction to the App Service Environments for more information.

App Service Environments are a great option if you want to isolate your functions, giving you more control over your instances.

You can get more details on how to use App Service Environments by checking out the Introduction to the App Service Environments.

Function App Duration is a crucial aspect of Azure Function Plans. The default execution time limit for the consumption-based pricing plan is five minutes, although users can increase this to 10 minutes.

You can adjust the default execution time limit by modifying the functionTimeout property in the host.json project file. This property applies specifically to function executions.

There are specific plans with default and maximum values for timeout durations: PlanDefaultMaximumFlex Consumption plan30UnboundedPremium plan30UnboundedDedicated plan30UnboundedContainer Apps30UnboundedConsumption plan510

The maximum amount of time that an HTTP triggered function can take to respond to a request is 230 seconds due to the default idle timeout of Azure Load Balancer. For longer processing times, consider using the Durable Functions async pattern or defer the actual work and return an immediate response.

The timeout duration for functions in a function app is defined by the functionTimeout property in the host.json project file.

See what others are reading: Azure Savings Plans

Scaling and performance are crucial considerations when choosing an Azure Function plan. You can manually scale out by adding more VM instances, or enable autoscale, though autoscale will be slower than the elastic scale of the Premium plan.

To scale up, you can choose a different App Service plan. The Premium plan, for example, offers elastic scale, which can automatically scale your app from zero to wherever your needs take you, and back.

The scaling behaviors of the various hosting plans are compared in the following table:

Some additional limits to keep in mind include a limit of 500 instances per subscription per hour for Linux apps on a Consumption plan, and a maximum of 1000 instances for Container Apps.

Broaden your view: Azure Function vs Logic App

Scaling is a crucial aspect of ensuring your Azure Functions app performs optimally. You can manually scale out by adding more VM instances, or enable autoscale, though autoscale will be slower than the elastic scale of the Premium plan.

Azure Functions also offers a pay-per-use model, allowing you to scale up or down based on your needs. With the Consumption plan, you can scale out automatically, even during periods of high load, with Functions infrastructure scaling CPU and memory resources by adding more instances of the Functions host.

You can also scale up by choosing a different App Service plan, such as the Premium plan, which allows for event-driven scaling and can scale out automatically even during periods of high load. The Premium plan can scale to 100 instances in some regions.

In terms of limits, there's currently a limit of 500 instances per subscription per hour for Linux apps on a Consumption plan. For specific limits for the various App Service plan options, see the App Service plan limits.

The following table compares the scaling behaviors of the various hosting plans:

Keep in mind that during scale-out, there's currently a limit of 500 instances per subscription per hour for Linux apps on a Consumption plan.

Cold start behavior can be a significant issue for applications that experience sudden spikes in traffic.

A Flex Consumption plan can help reduce the delay when provisioning new instances by supporting always ready instances.

The Premium plan takes it a step further by allowing you to maintain one or more perpetually warm instances, avoiding cold starts altogether.

In a Dedicated plan, the Functions host can run continuously on a prescribed number of instances, making cold start a non-issue.

Container Apps can also mitigate cold start behavior by setting the minimum number of replicas to one or more, ensuring the host process runs continuously.

Here's a breakdown of how different plans handle cold start behavior: